Patients with word retrieval defects for entities from varied conceptual categories or for actions, are still capable of retrieving many lexical items. The question thus arises as to which neural structures permit this "residual" word retrieval, a question that is important both to further our understanding of the neural basis of language processing, and to elucidate mechanisms of brain plasticity in the recovery from brain injury. In this project we will test the following three alternative hypotheses: (1) that residual word retrieval depends on the activation of regions in the right hemisphere, homologous to the ones activated in the left hemisphere in normal English speaking subjects; (2) that the residual capacity correlates with activation in the immediate anatomical surround of the damaged area; or (3) that the residual word retrieval depends on both activation of the right hemisphere homologues and the surround of the damaged area. We will study 40 brain damaged subjects in four PET experiments, involving the retrieval of words denoting concrete entities, as well as words denoting actions. Evidence from this project will be of great value to understand the process of recovery from brain injury in patients with English aphasia, as well as those with ASL aphasia, and to design effective rehabilitation strategies.